HFI design, qualification, and performance

The inversion of HFI data requires that one knows how the instrument selects photons, how these photons are transformed in data transmitted by telemetry and what spurious signals are added in this process.

HFI high level description and Architecture

(Lamarre/Pajot) Should be short, understandable and point through links to the relevant sections and papers.

Cryogenics

(F .Pajot)

Dilution

(including PIDs)

The HFI 3He-4He dilution cooler produces temperatures of 0.1 K for the bolometers through the dilution of 3He into 4He and 1.4 K through JT expansion of the 3He and 4He mixture. The dilution cooler is described in detail in [Planck early results. II. The thermal performance of Planck, 2.3.3. Dilution cooler].

The dilution was operated with flows set to the minimum available value, and provided a total lifetime of 30.5 months, exceeding the nominal lifetime of 16 months by 14.5 months. The dilution stage was stabilized by a PID control with a power comprised between 20 and 30 nW providing a temperature near 101 mK. The bolometer plate was stabilized at 102.8 mK with a PID power around 5 nW [fig. 100mK_stability.png].

(here a few lines of 100 mK boloplate stability)

Detailed of the in-flight performance of the dilution cooler can be found in [Planck early results. II. The thermal performance of Planck, 4.4. Dilution cooler]

4K J-T cooler

(including PIDs for details links to the early cryogenic paper (need to add data ?))

The HFI 4K J-T cooler produces a temperature of 4K for the HFI 4K stage and optics and the precooling of the dilution gases. Full description of the 4K cooler can be found in [Planck early results. II. The thermal performance of Planck, 2.3.2. 4He-JT cooler].

The 4K cooler was operated without interruption during all the survey phase of the mission. It is still in operation as it also provides the cooling of the optical reference loads of the LFI. The 4K PID stabilizing the temperature of the HFI optics is regulated at 4.81 K using a power around 1.8 mW [fig. 4K -A VENIR-].

(here a few lines of 4K stability, including compressors operation)

Details on the in-flight performance of the dilution cooler can be found in [Planck early results. II. The thermal performance of Planck, 4.3. 4He-JT cooler]

Cold optics

(Lamarre)

Horns,lenses

links to Peter's paper

filters, band

Includes Locke's very detailed document.

Detection chain

(Francesco Piacentini)

Time response.

The HFI bolometers and readout electronics have a finite response time to changes in incident optical power. The bolometers are thermal detectors of radiation whose response time is determined by the thermal circuit defined by the heat capacity of the detector and thermal conductivity.

Due to Planck's nearly constant scan rate, the time response is degenerate with the optical beam. However, because of the long time scale effects present in the time response, the time response is deconvolved from the data in the processing of the HFI data (see TOI processing).

The time response of the HFI bolometers and readout electronics is modeled as a Fourier domain transfer function (called the LFER4 model) consisting of the product of an bolometer thermal response [math]F(\omega)[/math] and an electronics response [math]H'(\omega)[/math].

[math]\label{LFER4def}TF^{LFER4}(\omega) = F(\omega) H'(\omega)[/math]

LFER4 model

If we write the input signal (power) on a bolometer as [math]\label{bol_in} s_0(t)=e^{i\omega t} [/math] the bolometer physical impedance can be written as: [math]\label{bol_out} s(t)=e^{i\omega t}F(\omega) [/math] where [math]\omega[/math] is the angular frequency of the signal and [math]F(\omega)[/math] is the complex intrinsic bolometer transfer function. For HFI the bolometer transfer function is modelled as the sum of 4 single pole low pass filters: [math]\label{bol_tf} F(\omega) = \sum_{i=0,4} \frac{a_i}{1 + i\omega\tau_i} [/math] The modulation of the signal is done with a square wave, written here as a composition of sine waves of decreasing amplitude: [math]\label{sigmod} s'(t)=e^{i\omega t}F(\omega)\sum_{k=0}^{\infty} \frac{e^{i\omega_r(2k+1)t}-e^{-i\omega_r(2k+1)t}}{2i(2k+1)} [/math] where we have used the Euler relation [math]\sin x=(e^{ix}-e^{-ix})/2i[/math] and [math]\omega_r[/math] is the angular frequency of the square wave. The modulation frequency is [math]f_{mod} = \omega_r/2\pi[/math] and was set to [math]f_{mod} = 90.18759 [/math]Hz in flight. This signal is then filtered by the complex electronic transfer function [math]H(\omega)[/math]. Setting: [math]\omega_k^+=\omega+(2k+1)\omega_r[/math] [math]\omega_k^-=\omega-(2k+1)\omega_r[/math] we have: [math]\label{sigele} \Sigma(t)=\sum_{k=0}^\infty\frac{F(\omega)}{2i(2k+1)}\left[H(\omega_k^+)e^{i\omega_k^+t}-H(\omega_k^-)e^{i\omega_k^-t}\right] [/math] This signal is then sampled at high frequency ([math]2 f_{mod} NS[/math]). [math]NS[/math] is one of the parameters of the HFI electronics and corresponds to the number of high frequency samples in each modulation semi-period. In order to obtain an output signal sampled every [math]\pi/\omega_r[/math] seconds, we must integrate on a semiperiod, as done in the HFI readout. To also include a time shift [math]\Delta t[/math], the integral is calculated between [math]n\pi/\omega_r+\Delta t[/math] and [math](n+1)\pi/\omega_r+\Delta t[/math] (with [math]T=2 \pi/\omega_r[/math] period of the modulation). The time shift [math]\Delta t[/math] is encoded in the HFI electronics by the parameter [math]S_{phase}[/math], with the relation [math]\Delta t = S_{phase}/NS/f_{mod} [/math].

After integration, the [math]n[/math]-sample of a bolometer can be written as [math]\label{eqn:output} Y(t_n) = (-1)^n F(\omega) H'(\omega) e^{i t_n \omega} [/math] where [math]\label{tfele} H'(\omega) = \frac 12 \sum_{k=0}^\infty e^{-i(\frac{\pi\omega}{2\omega_r}+\omega\Delta t)} \Bigg[ \frac{H(\omega_k^+)e^{i\omega_k^+ \Delta t}}{(2k+1)\omega_k^+} \left(1-e^{\frac{i\omega_k^+\pi}{\omega_r}}\right) \\ - \frac{H(\omega_k^-)e^{i\omega_k^- \Delta t}}{(2k+1)\omega_k^-} \left(1-e^{\frac{i\omega_k^-\pi}{\omega_r}}\right) \Bigg] [/math]

The output signal in equation eqn:output can be demodulated (thus removing the [math](-1)^n[/math]) and compared to the input signal in equation bol_in. The overall transfer function is composed of the bolometer transfer function and the effective electronics transfer function, [math]H'(\omega)[/math]: [math]TF(\omega) = F(\omega) H'(\omega) [/math]

The shape of [math]H(\omega)[/math] is obtained combining low and high-pass filters with Sallen Key topologies (with their respective time constants) and accounting also for the stray capacitance low pass filter given by the bolometer impedance combined with the stray capacitance of the cables. The sequence of filters that define the electronic band-pass function [math]H(\omega) = h_0*h_1*h_2*h_3*h_4*h_{5}[/math] are listed in table table:readout_electronics_filters.

Parameters of LFER4 model

The LFER4 model has are a total of 10 parameters([math]A_1[/math],[math]A_2[/math],[math]A_3[/math],[math]A_4[/math],[math]\tau_1[/math],[math]\tau_2[/math],[math]\tau_3[/math],[math]\tau_4[/math],[math]S_{phase}[/math],[math]\tau_{stray}[/math]) 9 of which are independent, for each bolometer. The free parameters of the LFER4 model are determined using in-flight data in the following ways:

• [math]S_{phase}[/math] is fixed at the value of the REU setting.
• [math]\tau_{stray}[/math] is measured during the QEC test during CPV.
• [math]A_1[/math], [math]\tau_1[/math], [math]A_2[/math], [math]\tau_2[/math] are fit forcing the compactness of the scanning beam.
• [math]A_3[/math], [math]\tau_3[/math], [math]A_{4}[/math] [math]\tau_4[/math] are fit by forcing agreement of survey 2 and survey 1 maps.
• The overall normalization of the LFER4 model is forced to be 1.0 at the signal frequency of the dipole.

The details of determining the model parameters are given in (reference P03c paper) and the best-fit parameters listed here in table table:LFER4pars.

HFI electronics filter sequence

HFI electronics filter sequence. We define $s = i \omega$

Filter	Parameters	Function
0. Stray capacitance low pass filter	[math]\tau_{stray}= R_{bolo} C_{stray}[/math]	[math]h_0 = \frac{1}{1.0+\tau_{stray}*s}[/math]
1. Low pass filter	[math]R_1=1[/math]k[math]\Omega[/math] [math]C_1=100[/math]nF	[math]h_1 = \frac{2.0+R_1*C_1*s}{2.0*(1.0+R_1*C_1*s)}[/math]
2. Sallen Key high pass filter	[math]R_2=51[/math]k[math]\Omega[/math] [math]C_2=1\mu[/math]	[math]h_2= \frac{(R_2*C_2*s)^2}{(1.0+R_2*C_2*s)^2}[/math]3
3. Sign reverse with gain		[math]h_3=-5.1[/math]
4. Single pole low pass filter with gain	[math]R_4=10[/math]k[math]\Omega[/math] [math]C_4=10[/math]nF	[math]h_4= \frac{1.5}{1.0+R_4*C_4*s}[/math]
5. Single pole high pass filter coupled to a Sallen Key low pass filter	[math]R_9=18.7[/math]k[math]\Omega[/math] [math]R_{12}=37.4[/math]k[math]\Omega[/math] [math]C=10.0[/math]nF [math]R_{78}=510[/math]k[math]\Omega[/math] [math]C_{18}=1.0\mu[/math]F [math]K_3 = R_9^2*R_{78}*R_{12}^2*C^2*C_{18}[/math] [math]K_2 = R_9*R_{12}^2*R_{78}*C^2+R_{9}^2*R_{12}^2*C^2+R_9*R_{12}^2*R_{78}*C_{18}*C[/math] [math]K_1 =R_9*R_{12}^2*C+R_{12}*R_{78}*R_9*C_{18}[/math]	[math]h_{5} = \frac{2.0*R_{12}*R_9*R_{78}*C_{18}*s}{s^3*K_3 + s^2*K_2+ s*K_1 + R_{12}*R_9 } [/math]

Parameters for LFER4 model.

Bolometer	[math]A_1[/math]	[math]\tau_1[/math] (s)	[math]A_2[/math]	[math]\tau_2[/math] (s)	[math]A_3[/math]	[math]\tau_3[/math] (s)	[math]A_4[/math]	[math]\tau_4[/math] (s)	[math]\tau_{stray}[/math] (s)	[math]S_{phase}[/math]
100-1a	0.392	0.01	0.534	0.0209	0.0656	0.0513	0.00833	0.572	0.00159	0.00139
100-1b	0.484	0.0103	0.463	0.0192	0.0451	0.0714	0.00808	0.594	0.00149	0.00139
100-2a	0.474	0.00684	0.421	0.0136	0.0942	0.0376	0.0106	0.346	0.00132	0.00125
100-2b	0.126	0.00584	0.717	0.0151	0.142	0.0351	0.0145	0.293	0.00138	0.00125
100-3a	0.744	0.00539	0.223	0.0147	0.0262	0.0586	0.00636	0.907	0.00142	0.00125
100-3b	0.608	0.00548	0.352	0.0155	0.0321	0.0636	0.00821	0.504	0.00166	0.00125
100-4a	0.411	0.0082	0.514	0.0178	0.0581	0.0579	0.0168	0.37	0.00125	0.00125
100-4b	0.687	0.0113	0.282	0.0243	0.0218	0.062	0.00875	0.431	0.00138	0.00139
143-1a	0.817	0.00447	0.144	0.0121	0.0293	0.0387	0.0101	0.472	0.00142	0.00125
143-1b	0.49	0.00472	0.333	0.0156	0.134	0.0481	0.0435	0.27	0.00149	0.00125
143-2a	0.909	0.0047	0.0763	0.017	0.00634	0.1	0.00871	0.363	0.00148	0.00125
143-2b	0.912	0.00524	0.0509	0.0167	0.0244	0.0265	0.0123	0.295	0.00146	0.00125
143-3a	0.681	0.00419	0.273	0.00956	0.0345	0.0348	0.0115	0.317	0.00145	0.00125
143-3b	0.82	0.00448	0.131	0.0132	0.0354	0.0351	0.0133	0.283	0.00161	0.000832
143-4a	0.914	0.00569	0.072	0.0189	0.00602	0.0482	0.00756	0.225	0.00159	0.00125
143-4b	0.428	0.00606	0.508	0.00606	0.0554	0.0227	0.00882	0.084	0.00182	0.00125
143-5	0.491	0.00664	0.397	0.00664	0.0962	0.0264	0.0156	0.336	0.00202	0.00139
143-6	0.518	0.00551	0.409	0.00551	0.0614	0.0266	0.0116	0.314	0.00153	0.00111
143-7	0.414	0.00543	0.562	0.00543	0.0185	0.0449	0.00545	0.314	0.00186	0.00139
217-5a	0.905	0.00669	0.0797	0.0216	0.00585	0.0658	0.00986	0.342	0.00157	0.00111
217-5b	0.925	0.00576	0.061	0.018	0.00513	0.0656	0.0094	0.287	0.00187	0.00125
217-6a	0.844	0.00645	0.0675	0.0197	0.0737	0.0316	0.0147	0.297	0.00154	0.00125
217-6b	0.284	0.00623	0.666	0.00623	0.0384	0.024	0.0117	0.15	0.00146	0.00111
217-7a	0.343	0.00548	0.574	0.00548	0.0717	0.023	0.0107	0.32	0.00152	0.00139
217-7b	0.846	0.00507	0.127	0.0144	0.0131	0.0479	0.0133	0.311	0.00151	0.00139
217-8a	0.496	0.00722	0.439	0.00722	0.0521	0.0325	0.0128	0.382	0.00179	0.00111
217-8b	0.512	0.00703	0.41	0.00703	0.0639	0.0272	0.0139	0.232	0.00173	0.00125
217-1	0.0136	0.00346	0.956	0.00346	0.0271	0.0233	0.00359	1.98	0.00159	0.00111
217-2	0.978	0.00352	0.014	0.0261	0.00614	0.042	0.00194	0.686	0.0016	0.00125
217-3	0.932	0.00355	0.0336	0.00355	0.0292	0.0324	0.00491	0.279	0.00174	0.00125
217-4	0.658	0.00135	0.32	0.00555	0.0174	0.0268	0.00424	0.473	0.00171	0.00111
353-3a	0.554	0.00704	0.36	0.00704	0.0699	0.0305	0.0163	0.344	0.0017	0.00125
353-3b	0.219	0.00268	0.671	0.00695	0.0977	0.0238	0.0119	0.289	0.00157	0.00111
353-4a	0.768	0.00473	0.198	0.00993	0.0283	0.0505	0.00628	0.536	0.00181	0.00125
353-4b	0.684	0.00454	0.224	0.0108	0.0774	0.08	0.0149	0.267	0.00166	0.00111
353-5a	0.767	0.00596	0.159	0.0124	0.0628	0.0303	0.0109	0.357	0.00156	0.00111
353-5b	0.832	0.00619	0.126	0.0111	0.0324	0.035	0.0096	0.397	0.00166	0.00111
353-6a	0.0487	0.00176	0.855	0.006	0.0856	0.0216	0.0105	0.222	0.00199	0.00125
353-6b	0.829	0.00561	0.127	0.00561	0.0373	0.0252	0.00696	0.36	0.00228	0.00111
353-1	0.41	0.000743	0.502	0.00422	0.0811	0.0177	0.0063	0.329	0.00132	0.00097
353-2	0.747	0.00309	0.225	0.00726	0.0252	0.0447	0.00267	0.513	0.00154	0.00097
353-7	0.448	0.0009	0.537	0.0041	0.0122	0.0273	0.00346	0.433	0.00178	0.00125
353-8	0.718	0.00223	0.261	0.00608	0.0165	0.038	0.00408	0.268	0.00177	0.00111
545-1	0.991	0.00293	0.00743	0.026	0.00139	2.6	0	0	0.00216	0.00111
545-2	0.985	0.00277	0.0128	0.024	0.00246	2.8	0	0	0.00187	0.00097
545-4	0.972	0.003	0.0277	0.025	0.000777	2.5	0	0	0.00222	0.00111
857-1	0.974	0.00338	0.0229	0.025	0.00349	2.2	0	0	0.00176	0.00111
857-2	0.84	0.00148	0.158	0.00656	0.00249	3.2	0	0	0.0022	0.00125
857-3	0.36	4.22e-05	0.627	0.0024	0.0111	0.017	0.002	1.9	0.00152	0.00126
857-4	0.278	0.0004	0.719	0.00392	0.00162	0.09	0.00152	0.8	0.00149	0.000558

System (Ken)

List of systematics

HFI validation is mostly modular. In other words, each part of the pipeline, e.g., timeline processing or mapmaking, has the results of its work validated at each step of the processing, looking specifically for known issues. In addition, we perform additional validation with an eye towards overall system integrity, by looking at generic differences between sets of maps, in which most problems will become apparent (whether known or not). Both these approaches are described below.

Expected systematics and tests (bottom-up approach)

Like all experiments, Planck-HFI had a number of specific issues that it needed to be tracked to verify that they were not compromising the data. While these are discussed in appropriate sections, here we gather them together to give brief summaries of the issues and refer the reader to the appropriate section for more details.

• Cosmic rays – unprotected by the atmosphere and more sensitive than previous bolometric experiments, HFI saw many more cosmic ray hits than its predecessors. These were detected, the worst parts of the data flagged as unusable, and "tails" were modelled and removed. This is described in the section on glitch statistics as well as in the 2013 HFI glitch removal paper.
• "Elephants" – cosmic rays also hit the HFI 100-mK stage and cause the temperature to vary, inducing small temperature and thus noise variations in the detectors. These elephants are removed with the rest of the thermal fluctuations, described directly below.
• Thermal fluctuations – HFI is an extremely stable instrument, but there are small thermal fluctuations. These are discussed in the timeline processing section on thermal decorrelation.
• Random telegraphic signal (RTS) or "popcorn noise" – some channels were occasionally affected by what seems to be a baseline that abruptly changes between two levels, which has been variously called popcorn noise or random telegraphic signal. These data are usually flagged. This is described in the section on noise stationarity.
• Jumps – similar to (but distinct from) popcorn noise, small jumps were occasionally found in the data streams. These jumps are usually corrected, as described in the section on jump corrections.
• 4-K cooler-induced EM noise – the 4-K cooler induced noise in the detectors with very specific frequency signatures, which can be filtered. This is described in the 2013 HFI DPC Paper; their stability is discussed in the section on 4-K cooler line stability.
• Compression – on-board compression is used to overcome our telemetry bandwidth limitations. This is explained in Planck-Early-IV.
• Noise correlations – correlations in noise between detectors seems to be negligible, except for two polarization-sensitive detectors in the same horn. This is discussed in the 2013 HFI Glitch removal paper.
• Pointing – the final pointing reconstruction for Planck is near the arcsecond level. This is discussed in the 2013 HFI DPC Paper.
• Focal plane geometry – the relative positions of different horns in the focal plane are reconstructed using planets. This is also discussed in the 2013 HFI DPC paper.
• Main beam – the main beams for HFI are discussed in the 2013 Beams and Transfer function paper.
• Ruze envelope – random imperfections, or dust on the mirrors, can mildly increase the size of the beam. This is discussed in the 2013 Beams and Transfer function paper.
• Dimpling – the mirror support structure causes a pattern of small imperfections in the beams, which generate small sidelobe responses outside the main beam. This is discussed in the the 2013 Beams and Transfer function paper.
• Far sidelobes – small amounts of light can sometimes hit the detectors from just above the primary or secondary mirrors, or even from reflections off the baffles. While small, when the Galactic centre is in the right position, this can be detected in the highest frequency channels, and so is removed from the data. This is discussed in the 2013 Beams and Transfer function paper and also in the 2013 Zodiacal emission paper.
• Planet fluxes – comparing the known flux densities of planets with the calibration on the CMB dipole is a useful check of calibration for the CMB channels, and is the primary calibration source for the submillimetre channels. This is done in the 2013 Mapmaking and Calibration paper.
• Point source fluxes – as with planet fluxes, we also compare fluxes of known, bright point sources with the CMB dipole calibration. This is done in the 2013 Mapmaking and Calibration paper.
• Time constants – the HFI bolometers do not react instantaneously to light; there are small time constants, discussed in the 2013 Beams and Transfer function paper.
• ADC correction – the HFI analogue-to-digital converters are not perfect, and are not used perfectly. Their effects on the calibration are discussed in the 2013 Mapmaking and Calibration paper.
• Bandpass – the transmission curves, or "bandpasses" have shown up in a number of places. This is discussed in the 2013 spectral response paper.

Summary

(Lamarre) here remind worse sytematics and point to DPC Summary of sucess and limitations. JML. Link to early HFI in flight perf.